Daniel Klein views the rise of government through Ngram

Here is the abstract:

In this very casual paper, I reproduce results from the Google Ngram Viewer. The main thrust is to show that around 1880 governmentalization of society and culture began to set in — a great transformation, as Karl Polanyi called it. But that great transformation came as a reaction to liberalism, the first great transformation. The Ngrams shown include liberty, constitutional liberty, faith, eternity, God, social gospel, college professors, psychology, economics, sociology, anthropology, political science, criminology, new liberalism, old liberalism, public school system, Pledge of Allegiance, income tax, government control, run the country, lead the country, lead the nation, national unity, priorities, social justice, equal opportunity, economic inequality, forced to work, living wage, social needs, our society, bundle of rights, property rights, capitalism, right-wing, left-wing, virtue, wisdom, prudence, benevolence, diligence, fortitude, propriety, ought, good conduct, bad conduct, good works, evil, sentiments, impartial, objective, subjective, normative, values, preferences, beliefs, and information.

The paper is here.  Here is one example:


What are the units for the Y-axis?

Percent of all published works in which that phrase appeared, I think.

See page 1 of the paper - it is the percentage of N-grams (2-grams in the "economic inequality" example) that are exactly that N-gram.

I think I am mistaken here - "economic inequality" is actually a 19-gram.

Also, the Democrats are onto something (Hail Satan * 666 vs. George Bush): http://books.google.com/ngrams/graph?content=Hail+Satan*666%2CGeorge+Bush&year_start=1920&year_end=2000&corpus=15&smoothing=3&share=

It's a two-gram, N is the word count, not the letter count.

If you plot an Ngram for just the word "economic" it looks almost identical with a rise after 1880. Is his effect spurious based on just when "economics" entered the popular lexicon?


In a word, yes.

See this Ngram comparing the use of the phrase "economic inequality" with the use of the phrase, "the social question," the term by which the problem of economic inequality was addressed basically immediately prior.


Actually you can do even better:

See how 'inequality' and 'economic inequality' start to track together after 1955.

Nice. I didn't know you could so easily weight ngrams by just adding a multiplier like that. Cool trick.

Ha, I just randomly amused myself. "Trick." "trick to hide the decline." Remember that whole ridiculous thing? 'He said trick, it must be intentional deception and trickery, to the outrage mobile!' Sorry, being random.

A lot of this seems to run parallel with the growth of the social sciences. I'm not certain that a growth in social sciences terminology portends much apart from a relative growth in the publishing output of the social sciences.

Ngrams is fun but it feels like the bluntest of objects with which to analyze here. I feel like I'd want specifically to analyze a certain corpus of writing, and not get pulled along by a changing mix in publishing trends (especially in areas where the incentives to publish may skew things, or Google Books/Scholar is particularly strong in terms of its content- academia.) You can say this points to a growth in the influence of academia, but I'd take a step back and say that we can only say it points to a growth in the publishing output of academia, which are surely two different things.

That said, here's a fun graph:

Hail Satan vs. Hail Hitler. ("Heil Hitler" is omitted because it dwarfs both and tracks very well with "Hail Hitler", and "Ave Sathanas" is omitted because it is an archaic form not even found at all in Ngrams.)

It would be interesting to see greater detail on the left, though the data are probably all a bit suspect.

Every major economic revolution since Industrial Revolution began has made some new class of “wealthy” even richer than the last (in present time) and increased income disparities. Why? Because every such revolution has increased value creation at a faster absolute pace. Therefore, it’s quite likely that the end of growing income disparity would also be the end of economic growth.

Our public policy goal should be facilitating a process that is meritocratic and which rewards value creation, not coercing society into a flat income distribution where almost everyone is poorer in absolute terms than they would have been.

Then shouldn't that public policy goal be to replace all taxes on economic activity (income, sales, capital-gains, etc.) with a single wealth tax i.e. tax on net-assets?

An excise tax on wealth? Wouldn't an excise tax on savings be more likely to contribute toward growth? I guess you could call that the Cypriot school.

Mmm...seed popcorn!

It's an interesting idea! I wonder if it's ever been tried in modern times. I'm not sure how it would work out -- greater incentives for income, smaller for savings, a bit of a reverse consumption tax?

Currently most people have most of their wealth in their homes, and they all pay property tax, so they're already burdened with a wealth tax. On top of this wealth tax, people are burdened with all sorts of taxes on economic activity - income taxes, sales taxes, capital-gains taxes, etc. People are being taxed for working, for engaging in commerce, for investing, etc. There is a basic justification for a wealth tax - the government's primary function is to protect your wealth. If your property includes, say, Florida coastline, you own it to the extent that the government enforces your property right and keeps people off of it and protects it from foreign invasion. If Cuba invades and seizes the Florida coastline, you don't own it anymore. But there's really no justification for all the other taxes on economic activity. There's no reason why people should be taxed for using their own property, such as their labor or capital, to generate return.

If it were tried, I imagine it would lead to an explosion in economic activity and growth, since an enormous tax burden would be lifted. Total wealth in the US was estimated at $118 trillion in 2008. Total federal spending in the US in 2008 was estimated at $2.9 trillion. A single wealth tax at 2.5% would have covered all the spending for that year. Furthermore, it would clarify in the minds of the citizenry the proper, limited role of government as the protector of property rights from domestic thieves and foreign military invasion. More people would realize that the primary function of government is to protect property rights and that to fulfill that function it does not need to spend 2.5% of total wealth annually to do so. It can do so with a lot less. US defense spending in 2008 was $600 billion. That spending could have been covered with a single wealth tax at 0.5%. But we know that even defense spending is bloated and exceeds what is necessary for the protection of property rights from domestic thieves and foreign military invasion. So the single wealth tax should be less than 0.5%.

A single wealth tax at 0.1% in 2008 would have furnished $118 billion in revenue for the government, more than enough for the police and military to defend US property.

With a 0.1% single wealth tax, a businessman with, say, $1 million in net assets, who gets a net profit of $100,000 for the year running his business, would have a tax burden of just $1,000. He would pay 0.1% on his net assets as a fee for the government protecting his net assets, but he would pay nothing on his net profit or anything else since the government has nothing to do with that. The government played no role in generating the profit for the business except for providing protection of the enterprise's assets. The only way it could help the business generate profit is by somehow hurting someone else via regulation, barriers to entry, etc.

I guess as an accountant I see the major problems as technical -- many assets are illiquid and hence difficult to price accurately. And it's hard to know who might have a stack of gold bars hidden in their basement going untaxed. That's probably the main reason income is the primary tax source -- it's easiest to identify.

Still, interesting notion. It does have some elegant features. Maybe it could be workable.

Wealth is easier to hide than income, which is a sorta proxy for wealth anyway.

Inflation is also a handy wealth tax, which runs about 2% per year.

I agree that wealth taxes are best- it's largely what we do.

What aprox. % of wealth is in corporate equity, financial assets, and real estate? Those are already registered with the state, and fairly easy to track, and are not easy to hide.

They are also the primary forms of capital that create profits. What if we just had a 1% wealth tax on those forms of wealth?

The idea of getting rid of other federal taxes and replacing it with a shallow, progressive wealth tax is appealing. If we all pooled our resources together for common defense and justice, wouldn't we pro-rate what we contributed by the assets

Reminds me of an intelligent librarian who told a few that the word "strike" in the 1880's was classified as a sub-heading of " criminal activity"! Sure, words condition us, and we condition words, but that's neither deterministic nor causal. Once again, just a bunch of words, in a more neutral sense than in my comment on some left wing imbecile's work.

The more people yak about liberty the less there is of it?

There has to be a bias -- perhaps controlled for as I didn't read the paper, just skimmed very quickly -- in that historical texts and usages have to be under represented in the search.

Are you assuming that the numerator of the Y axis changes with time but the denominator is aggragated over the whole database? That would be odd.

What is true is that older books related to the speech of the day differently for how current books relate to speech of our day. N-Gram can only tell us about the bookish worold.

It looks like a chart on the rise of communists/leftism. There's a dip into the 1950s when the U.S. had a reaction and started looking for communist infiltration. By the end of the 1960s they were taking over and are now running the show.

"It looks like a chart on the rise of communists/leftism."

Interesting theory. To me looks like the rise of a--turds. There's a dip in the 50s when a lot of Americans grew tired of a--turds after the war but there's been a steady rise ever since. Now a--turds are running the show.

There was a dip in the 50's because there was less economic inequality.

Does the rise in government correlate to the fall in crime and the improvement in health?

Tyler's forthcoming book "Average is Over" just hit Amazon.

Quality articles is the important to interest the people
to pay a quick visit the site, that's what this web site is providing.

My spouse and I stumbled over here from a different web address and thought I should check things out.

I like what I see so now i am following you. Look forward to looking into your web page yet again.

Comments for this post are closed